Dynamic spectrum frequency hopping communication method and device for target drone and ground station

Abstract

The invention relates to the technical field of unmanned aerial vehicle communication, and discloses a dynamic spectrum frequency hopping communication method and device for a target aircraft and a ground station, and the method comprises the steps: collecting state information; evaluating a signal quality indicator based on the power data; based on the total spectral range and the signal quality index, segmenting the total spectral range into N sub-bands to obtain a spectrum allocation result; based on the task priority and the communication quality index, carrying out main target machine election; each target machine sends an orthogonal detection sequence in a corresponding time slot to obtain channel state information; converting the channel state information into a frequency hopping key; generating a frequency hopping sequence based on the frequency hopping key; according to the invention, an adaptive frequency hopping technology based on a channel state is adopted, the channel quality is monitored in real time, and a dynamic frequency hopping sequence is generated, so that an interference frequency band is effectively avoided; through a time division multiplexing channel detection mechanism, signal overlapping is fundamentally avoided, and continuous and stable operation in a complex dynamic environment is ensured.

Description

Technical Field

[0001] The present invention relates to the field of unmanned aerial vehicle (UAV) communication technology, and more particularly to a dynamic spectrum hopping communication method and device between a target drone and a ground station. Background Art

[0002] With the increasing demand for modern training activities and equipment testing, unmanned target drone systems have become an important training equipment. In a typical target drone training scenario, multiple target drones are usually required to perform different flight missions simultaneously, including single-aircraft target simulation, formation flying, coordinated maneuvers, and other complex tasks. This requires a stable and secure communication link between the target drones and the ground station.

[0003] Currently, drone communication systems primarily utilize fixed frequencies or simple frequency hopping. Fixed-frequency communication is susceptible to interception and interference, while traditional frequency hopping technology requires pre-set hopping patterns, presenting numerous technical challenges when operating with multiple drones in coordination.

[0004] The existing Chinese patent with authorization announcement number CN113765541B discloses a UAV frequency hopping communication method, device, computer equipment and storage medium, and provides a frequency hopping communication scheme that dynamically generates a consensus frequency hopping sequence based on channel estimation technology. That is, without relying on a dedicated frequency hopping system, a frequency hopping sequence that does not require prior agreement can be dynamically generated between ordinary ground stations (or UAVs) and UAVs, thereby meeting the dynamic networking requirements between different UAVs and ground stations. It can enable the two UAVs to communicate based on the generated frequency hopping sequence and continuously switch frequencies according to time slices to communicate. It can also enable the UAVs that communicate by frequency hopping based on the dynamic generation of the frequency hopping sequence to greatly increase the confidentiality of their communications and prevent eavesdroppers from stealing information.

[0005] However, the existing technology is only applicable to point-to-point communication scenarios. When faced with one-to-many communication between multiple target aircraft and ground stations, when multiple target aircraft send channel detection signals to the ground station at the same time, signal overlap and mutual interference will occur, affecting the accurate calculation of channel state information. Summary of the Invention

[0006] The purpose of the present invention is to provide a dynamic spectrum hopping communication method and device between a target drone and a ground station in order to solve the above problems.

[0007] The present invention provides a dynamic spectrum hopping communication method between a target drone and a ground station, comprising the following steps:

[0008] Collect status information, including task priority and total spectrum range;

[0009] Monitor power data in real time and evaluate signal quality indicators based on the power data. Based on the total spectrum range and signal quality indicators, divide the total spectrum range into N sub-bands to obtain spectrum allocation results.

[0010] Calculate communication quality indicators based on power data, select the master target based on task priority and communication quality indicators, and establish a timing benchmark;

[0011] Based on the timing reference, the time axis is divided into N+1 non-overlapping time slots. Each target machine sends an orthogonal sounding sequence in the corresponding time slot to obtain channel state information.

[0012] The channel state information is converted into a frequency hopping key; and a frequency hopping sequence is generated based on the frequency hopping key.

[0013] Furthermore, the frequency hopping sequence includes a master frequency hopping sequence and a slave frequency hopping sequence, and the steps of generating the frequency hopping sequence include:

[0014] The master target machine uses the generated master frequency hopping key to perform a modulo operation on the size of the first frequency point set, and adds one to the remainder as the index value. The master target machine selects the frequency point at the corresponding position in the first frequency point set as the frequency hopping frequency at the current moment. The frequency hopping frequencies at multiple moments constitute the master frequency hopping sequence. The first frequency point set is the set of frequencies that meet the minimum interval constraint and the signal quality index is greater than the preset red threshold.

[0015] The slave target machine generates a slave frequency hopping key according to the received channel state information; processes the slave frequency hopping key through the SHA256 hash function; receives the master frequency hopping sequence broadcast by the master target machine; and applies an orthogonal mapping function to combine the hashed slave frequency hopping key and the master frequency hopping sequence to generate a slave frequency hopping sequence.

[0016] Furthermore, the frequency hopping key includes a master frequency hopping key and a slave frequency hopping key. The frequency hopping key is generated by using a multi-channel feature fusion method to extract key information of a path with a maximum modulus value of a complex gain coefficient in the channel state information. The key information includes the modulus value of the complex gain coefficient, the amplitude angle of the complex gain coefficient, and the time rate of change of the channel state. The key information is quantized, and the quantization results are connected in series to obtain a basic key. The basic key is shifted and XORed to obtain a mixed key. The mixed key is subjected to multiple shift and connection operations to obtain the frequency hopping key.

[0017] The modulus of the complex gain coefficient is the sum of the square of the real part and the square of the imaginary part of the complex gain coefficient, and then the square root of the sum is taken; the amplitude of the complex gain coefficient is the imaginary part of the complex gain coefficient divided by the real part, and the inverse tangent function of the result is obtained; the time rate of change of the channel state is obtained by the ratio of the modulus of the change in the complex gain coefficient at adjacent moments to the time interval, and the change in the complex gain coefficient at adjacent moments is obtained by subtracting the complex gain coefficient at the previous moment from the complex gain coefficient at the current moment.

[0018] Furthermore, the orthogonal mapping function converts the hashed slave frequency hopping key, the frequency hopping frequency at the current moment in the master frequency modulation sequence, the slave target machine number and the current timestamp into thirty-two bits and then performs an XOR operation to obtain a composite key value; a modulo operation is performed on the composite key value and the number of first frequency points in the slave target machine, and the remainder obtained is added by one as the basic frequency point index; and the frequency point at the corresponding position is selected from the first frequency point set of the corresponding slave target machine through the basic frequency point index as the frequency hopping frequency at the current moment.

[0019] Furthermore, the specific steps of dividing the time slots into N+1 non-overlapping time slots include:

[0020] All target drones and ground stations are time synchronized based on the timing reference;

[0021] Split the continuous time axis into periodic time frames;

[0022] Each time frame is divided into N+1 time slots of equal length. The first N time slots are allocated to N target drones, and the N+1th time slot is allocated to the ground station.

[0023] Furthermore, the sounding signal is a signal of an orthogonal sounding sequence propagated through a wireless channel. The orthogonal sounding sequence is a signal sequence with mutually orthogonal characteristics. The channel state information includes a channel impulse response, the number of multipath taps, and multipath propagation parameters. The channel impulse response is composed of complex gain coefficients of multiple multipath components. The multipath propagation parameters include the time delay, modulus attenuation, and amplitude offset of each path.

[0024] The detection signal is correlated with a known orthogonal detection sequence. The correlation operation is performed by integrating the conjugate of the detection signal and the orthogonal detection sequence at different time delay positions. The integration result is the channel impulse response corresponding to the corresponding time delay position.

[0025] The number of multipath taps is determined by a power threshold detection method. Multipath components exceeding a preset threshold are determined to be valid taps. The number of valid taps is the number of multipath taps.

[0026] The path delay is determined by the position of the channel impulse response in the time domain. The delay is equal to the time position of the peak minus the arrival time of the direct path. The modulus attenuation is calculated by the modulus of the peak. The argument offset is calculated by the argument of the peak.

[0027] Furthermore, the comprehensive score of each target drone is calculated by weighted summation, and the target drone with the highest comprehensive score is selected as the main target drone; the comprehensive score is the sum of the task priority item and the communication quality index item, wherein the task priority item is the normalized task priority multiplied by the preset task priority weight, the communication quality index item is the normalized communication quality index multiplied by the preset communication quality weight, and the communication quality is the ratio of the signal-to-noise ratio to the bit error factor; wherein the bit error factor is the distance factor multiplied by the bit error rate, the signal-to-noise ratio is the ratio of the useful signal power to the noise power, the distance factor is the propagation loss caused by the distance from the target drone to the ground station, and the bit error rate is the ratio of the number of received error bits to the total number of transmitted bits.

[0028] Furthermore, the steps of establishing a timing benchmark include:

[0029] Get GPS or BeiDou time from ground stations;

[0030] Generate multiple standard frequency components of different frequencies according to GPS or Beidou time, including 10MHz, 1MHz, and 100kHz;

[0031] Perform frequency domain superposition and synthesis of each standard frequency component to generate a timing reference signal;

[0032] Broadcast the timing reference signal to all target machines.

[0033] Furthermore, the power data includes signal power, noise power and interference power. The signal quality index is the ratio of signal power to the total noise. The total noise is the sum of noise power and interference power. The spectrum allocation result includes the sub-band allocated to each target aircraft and the first frequency point set in the corresponding sub-band.

[0034] The present invention provides a dynamic spectrum hopping communication device for a target drone and a ground station, which is used to store computer instructions and execute the aforementioned dynamic spectrum hopping communication method for a target drone and a ground station when the computer instructions are read. The device includes:

[0035] The collection module collects status information, including task priority and total spectrum range;

[0036] The allocation module monitors power data in real time and evaluates signal quality indicators based on the power data. Based on the total spectrum range and signal quality indicators, it divides the total spectrum range into N sub-bands to obtain the spectrum allocation results.

[0037] The timing module calculates the communication quality index based on the power data, selects the master target based on the task priority and the communication quality index, and establishes the timing benchmark;

[0038] The channel module divides the time axis into N+1 non-overlapping time slots based on the timing reference. Each target machine sends an orthogonal detection sequence in the corresponding time slot to obtain channel state information.

[0039] The frequency hopping module converts the channel state information into a frequency hopping key and generates a frequency hopping sequence based on the frequency hopping key.

[0040] The beneficial effects of the present invention are as follows: Through dynamic spectrum allocation and intelligent frequency hopping technology, the present invention achieves flexible sharing and optimized configuration of spectrum resources. It uses spectrum allocation optimization based on a genetic algorithm to dynamically adjust frequency band allocation based on real-time spectrum occupancy and task priority, thereby improving spectrum utilization. It also uses adaptive frequency hopping technology based on channel status to effectively avoid interfering frequency bands by monitoring channel quality in real time and generating dynamic frequency hopping sequences. The multi-frequency modulated composite timing reference signal has strong anti-interference properties, ensuring that even if some frequency components are interfered with, the other components can still maintain time synchronization.

[0041] Through the time-division multiplexing channel detection mechanism, the continuous time axis is divided into N+1 non-overlapping time slots, ensuring that each target machine sends a detection signal within a dedicated time slot, fundamentally avoiding signal overlap and ensuring the accuracy of channel state information calculation.

[0042] Through orthogonal frequency hopping sequence generation technology, it is ensured that the frequencies used by different target machines at any time are completely orthogonal; the method of combining SHA256 hash function and orthogonal mapping function is adopted to fundamentally eliminate the frequency conflict between multiple target machines; the present invention establishes a timing benchmark based on GPS or Beidou time, and adopts an adaptive broadband phase-locked loop to achieve precise synchronization under high dynamic conditions, thereby improving synchronization accuracy and ensuring the coordinated operation of various parts of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 It is an example flow chart of the dynamic spectrum hopping communication method between the target drone and the ground station of the present invention;

[0044] Figure 2 This is an example diagram of obtaining channel state information in the dynamic spectrum hopping communication method between the target aircraft and the ground station of the present invention;

[0045] Figure 3 This is an example of the frequency hopping sequence generation process of the dynamic spectrum hopping communication method between the target aircraft and the ground station of the present invention. Figure 1 ;

[0046] Figure 4 This is an example of the frequency hopping sequence generation process of the dynamic spectrum hopping communication method between the target aircraft and the ground station of the present invention. Figure 2 ;

[0047] Figure 5This is a module example diagram of the dynamic spectrum hopping communication device between the target aircraft and the ground station of the present invention. DETAILED DESCRIPTION

[0048] The subject matter described herein will now be discussed with reference to exemplary embodiments. It should be understood that these embodiments are discussed solely to enable those skilled in the art to better understand and implement the subject matter described herein, and that the functions and arrangements of the elements discussed may be varied without departing from the scope of this specification. Various examples may omit, substitute, or add various processes or components as needed. In addition, features described with respect to some examples may also be combined in other examples.

[0049] The dynamic spectrum hopping communication method and device between the target drone and the ground station include the following embodiments:

[0050] Example 1:

[0051] Dynamic spectrum hopping communication method between target drone and ground station, such as Figure 1 As shown, the following steps are included:

[0052] Step 101: Establish a basic data link connection between N target drones and the ground station in the training airspace through a standard data link protocol; collect status information, including target drone status information and spectrum environment information, to provide basic data for subsequent master-slave target drone election and spectrum allocation.

[0053] When the communication method is activated, a basic communication connection and a comprehensive perception of the working environment need to be established. First, within the training airspace, N target drones (N ≥ 2) establish a basic data link connection with the ground station. The number of target drones N is determined based on actual training needs and is usually in the range of 2-20 to balance training effectiveness and implementation complexity. The training airspace refers to a pre-demarcated aerial training area, usually a rectangular or circular area with a side length of 50-200 kilometers, ranging in altitude from the ground to 12,000 meters. It is demarcated by the training management department based on the type of training mission, safety requirements, and airspace control regulations.

[0054] The communication method uses a standard data link protocol to establish the initial connection, which includes three phases: device discovery, identity authentication, and connection verification. The advantage of using a standard data link protocol lies in its mature security mechanisms and high reliability, ensuring connection stability and security in complex electromagnetic environments. The standard data link protocol includes a complete protocol stack covering the physical, data link, network, and application layers, supporting functions such as encryption and authentication, error control, and traffic management. The ground station discovers available target drones through broadcasting. Each target drone is equipped with a unique identifier for identity authentication. After the connection is established, a communication quality test is performed. The ground station maintains a real-time target drone registry to monitor the connection status and communication quality of each target drone.

[0055] After successfully establishing a basic communication connection, it is necessary to collect status information to lay the foundation for subsequent coordinated communication. Status information includes target aircraft status information and spectrum environment information. The specific collection method is as follows:

[0056] Methods for collecting target drone status information include:

[0057] Position Coordinate Information: The precise three-dimensional position coordinates of each target drone, including longitude, latitude, and altitude, are acquired in real time using the GPS or BeiDou navigation system. GPS or BeiDou navigation systems are essential for their global coverage and high-precision positioning, providing an accurate position reference for subsequent Doppler effect calculations and channel predictions. Coordinate accuracy must reach the meter level, with an update frequency of 10-100 times per second. Airborne navigation equipment is equipped with a differential GPS or BeiDou receiver, which uses ground reference station corrections to enhance positioning accuracy.

[0058] Motion status information: The real-time motion status of the target drone is calculated through inertial navigation systems, GPS, or Beidou differential positioning, including velocity components and acceleration information in three directions. The acquisition of motion status data is crucial for predicting future positions and Doppler frequency shifts, as these parameters directly affect the changing trends of channel characteristics. The inertial navigation system is an autonomous navigation system based on gyroscopes and accelerometers. It measures the angular velocity and acceleration of the carrier to infer position and attitude. It has the advantages of being independent of external signals and having strong anti-interference capabilities, but it has the problem of error accumulation and requires regular correction through GPS or Beidou signals. Motion status data is processed through multi-sensor fusion, combining gyroscopes, accelerometers, and GPS or Beidou data to provide accurate dynamic information.

[0059] Mission priority information: Assign a mission priority value based on the importance and urgency of the target drone's mission. The value range is usually 1-10, and the higher the value, the higher the priority. The setting of mission priority is the basis for achieving differentiated service quality, ensuring that important missions obtain better communication resources and higher service quality. The mission importance refers to the degree of impact of the mission on the overall training objectives. For example, the priority of the simulation mission of the main attack target is 9-10, the priority of the auxiliary interference mission is 5-7, and the priority of the reconnaissance and surveillance mission is 3-5. The urgency refers to the time sensitivity of the mission execution. For example, the threat avoidance mission that requires an immediate response has a priority of 8-10, and the priority of the scheduled cruise mission is 2-4. The mission priority is dynamically adjusted by the mission control center according to the training plan, taking into account factors such as the mission type, execution difficulty and time requirements.

[0060] Methods for collecting spectrum environment information include:

[0061] Total spectrum range information: Spectrum monitoring equipment determines the frequency range within which communication methods operate, typically covering the L-band to Ku-band (1 GHz-18 GHz). This range is dynamically adjusted based on usage licenses and environmental conditions. Accurately determining the spectrum range is essential for preventing illegal occupation and interference, and also provides the boundaries of available resources for subsequent intelligent spectrum allocation. The L-band, which covers the 1 GHz-2 GHz frequency range, is primarily used for long-distance communications and navigation; the S-band, which covers the 2 GHz-4 GHz frequency range, is primarily used for radar and satellite communications; the C-band, which covers the 4 GHz-8 GHz frequency range, is primarily used for satellite communications and microwave relays; the X-band, which covers the 8 GHz-12 GHz frequency range, is primarily used for radar and high-speed data transmission; and the Ku-band, which covers the 12 GHz-18 GHz frequency range, is primarily used for satellite communications and high-resolution radar. The license specifies conditions such as the frequency range, transmit power limit, operating time, and geographic scope. Spectrum management devices monitor the status of spectrum licenses in real time to ensure compliance.

[0062] Frequency Spacing Information: The minimum spacing between adjacent first frequency bands is set based on the signal bandwidth and interference tolerance, typically 1.2-2 times the signal bandwidth, to ensure that adjacent signals do not interfere with each other. Proper frequency spacing can effectively prevent adjacent channel interference and improve spectrum efficiency. The signal bandwidth refers to the frequency range occupied by the communication signal; the interference tolerance is the ratio of adjacent channel interference power to the desired signal power that the receiver can tolerate, typically requiring adjacent channel interference to be 20-40 decibels lower than the desired signal. Spacing setting takes into account factors such as transmitter frequency stability, receiver selectivity, and the propagation environment.

[0063] Real-time spectrum occupancy information: Continuously monitors the current occupancy of each frequency point, identifies external interference sources and internal signal conflicts, and establishes a dynamic spectrum occupancy database. Real-time monitoring is essential due to the dynamic nature of the spectrum environment; only by understanding the real-time status can accurate spectrum allocation decisions be made. External interference sources include civilian communications equipment, industrial equipment radiation, radar systems, and other communications systems, which are classified using signal feature recognition algorithms. Internal signal conflicts refer to overlapping frequency usage between different users within the same communications system, which are identified through power detection and time domain analysis. The monitoring device uses a wideband receiver and spectrum analyzer, updating the spectrum status every 100 milliseconds.

[0064] After completing the status information collection, the communication method obtains the basic data for intelligent spectrum management, providing a decision basis for the next step of dynamic spectrum allocation.

[0065] Step 102: monitor power data in real time and evaluate signal quality indicators based on the power data; divide the total spectrum range into N sub-bands based on the total spectrum range and the signal quality indicators to obtain spectrum allocation results.

[0066] Specific implementation methods include:

[0067] The ground station's broadband spectrum analyzer simultaneously monitors spectrum status along three key dimensions, generating power data. The first dimension is signal power monitoring, using a square-law detector to measure the desired signal power strength at each frequency point in real time. The second dimension is noise power assessment, using power statistics during idle periods to determine the background noise level. The third dimension is interference power identification, using a signal signature analysis algorithm to distinguish and quantify the power of various interfering signals. A square-law detector is a nonlinear device whose output voltage is proportional to the input signal power, typically implemented using a diode or field-effect transistor. Idle periods are periods without desired signal transmission, and pure noise power measurements are obtained through timed sampling with a sampling interval of 10 milliseconds and a duration of 1 millisecond. The signal signature analysis algorithm, which includes power spectral density analysis, modulation identification, and time-frequency analysis, can distinguish between different types of interference, including continuous wave interference, pulsed interference, and modulated signal interference. The spectrum status monitoring method represents the status of each frequency point at a specific moment as a three-dimensional data combination consisting of signal power, noise power, and interference power. This representation comprehensively reflects spectrum usage and communication quality potential, providing an accurate basis for subsequent intelligent allocation decisions.

[0068] Signal quality indicators are used to quantify the availability of each frequency. The signal quality indicator is calculated by taking signal power as the numerator and the sum of noise power and interference power as the denominator. The resulting ratio reflects the signal quality indicator for that frequency. Higher signal quality indicators indicate better signal quality and are suitable for critical communications. Lower signal quality indicators indicate severe interference and should be avoided or protective measures implemented. Spectrum utilization is defined as the ratio of the actual spectrum bandwidth used to the total available spectrum bandwidth and measures the efficiency of spectrum resource utilization. The desired signal power refers to the power of the desired received communication signal, expressed in decibel milliwatts (dBm), with a typical range of -100 to -30 dBm. Noise power refers to the sum of thermal noise and device noise, calculated as 10 times the common logarithm of the product of the Boltzmann constant, absolute temperature, and signal bandwidth. The absolute temperature is typically 290 Kelvin, and the signal bandwidth is measured in Hertz. Interference power refers to the power of unwanted signals generated by external interference sources and internal signal conflicts. It is separated from the total received power using a signal recognition algorithm.

[0069] To improve assessment accuracy, the processing device uses a digital signal processor for real-time calculations, combining sliding average and exponentially weighted averaging to smooth the measurement results. Frequency points are classified into three categories based on the signal quality index calculation results, resulting in a frequency quality assessment result. High-quality frequencies are marked green, indicating excellent communication quality; average frequencies are marked yellow, indicating usable but requiring attention; and low-quality frequencies are marked red, indicating they should not be used. The sliding average is the arithmetic mean of the most recent M sample values, with M typically ranging from 10 to 50, which smooths out random fluctuations. The exponentially weighted average assigns decreasing weight to historical data. The calculation method is: the current average value is equal to the smoothing factor multiplied by the current measurement value plus one minus the smoothing factor multiplied by the previous average value. The smoothing factor is typically 0.1 to 0.3. The thresholds for the three-category classification are set as follows: a green frequency indicates a signal quality index greater than the green threshold (default value 10 decibels); a yellow frequency indicates a signal quality index between the green and red thresholds; and a red frequency indicates a signal quality index less than the red threshold (default value 0 decibels).

[0070] Based on signal quality indicators and real-time spectrum occupancy information, a genetic algorithm is used to optimize frequency allocation, achieving intelligent spectrum resource allocation. A genetic algorithm is a global optimization algorithm based on the principles of biological evolution. It simulates the processes of natural selection, crossover, and mutation to search for the optimal solution within a complex solution space. The advantage of using a genetic algorithm lies in its strong global search capabilities and ability to handle multi-objective optimization problems, making it suitable for complex combinatorial optimization problems such as spectrum allocation.

[0071] The genetic algorithm's optimization objective function comprehensively considers three key metrics: maximizing overall efficiency, optimizing load balancing, and minimizing interference. Maximizing overall efficiency refers to maximizing the system's overall communication capacity and service quality within given spectrum resource constraints. Optimizing load balancing ensures that the load distribution across frequency bands is as even as possible, avoiding overloading some bands while leaving others idle. Minimizing interference improves communication quality by reducing mutual interference through reasonable frequency allocation. These three objectives are weighted and summed to form a comprehensive objective function, with the weighting coefficients adjusted based on actual application requirements.

[0072] Based on the optimization objective, the genetic algorithm divides the total spectrum into N sub-bands, each of which is assigned to a specific target drone. Band segmentation employs an adaptive segmentation strategy. Based on individual frequency quality assessments and mission priority information, consecutive frequencies of similar quality are grouped into the same band, ensuring relatively uniform frequency quality within each band. Target drones assigned to high-priority missions are prioritized for high-quality sub-bands, while those assigned to low-priority missions are assigned to the remaining available sub-bands.

[0073] Sub-band allocation is achieved through chromosome encoding in a genetic algorithm. The chromosome encoding uses a one-dimensional array structure, with each element representing the sub-band index assigned to the corresponding target drone. The fitness function comprehensively considers three factors: frequency quality, task priority matching, and interference level, calculating the fitness value of each chromosome through a weighted summation. Genetic operations include selection, crossover, and mutation. The selection operation uses a roulette wheel selection method to select outstanding individuals based on their fitness values. The crossover operation uses single-point crossover to generate a new sub-band allocation scheme. The mutation operation randomly changes the sub-band allocation of some targets to increase population diversity.

[0074] The genetic algorithm parameters were set as follows: population size of 50-100 individuals, evolutionary generations of 100-200 generations, crossover probability of 0.7-0.9, and mutation probability of 0.01-0.1. The algorithm convergence criterion was that the optimal fitness value changed by less than 0.1% for 10 consecutive generations, or that the maximum number of evolutionary generations was reached.

[0075] Based on the sub-band allocation, a first frequency set is established for each target drone within its dedicated sub-band. This process involves first extracting all frequencies within the sub-band assigned to the target drone; then, based on frequency quality classification, removing low-quality frequencies marked in red; and finally, ensuring that adjacent selected frequencies meet the minimum spacing constraint based on frequency spacing requirements, forming the final first frequency set for the target drone.

[0076] The minimum interval constraint is determined based on the frequency spacing information collected in step 101. Specifically, the minimum frequency spacing between adjacent first frequency points is 1.2-2 times the signal bandwidth. The signal bandwidth is determined by the type of communication service. The digital data transmission bandwidth is typically 10-100kHz, and the high-definition video transmission bandwidth is 1-10MHz. For example, when the signal bandwidth is 50kHz, the minimum interval constraint is 60-100kHz; when the signal bandwidth is 2MHz, the minimum interval constraint is 2.4-4MHz. This constraint is designed to prevent adjacent channel interference, ensuring that signals on adjacent frequency points do not affect each other, thereby improving spectrum utilization efficiency and communication quality.

[0077] The minimum interval constraint is implemented as follows: All first-order frequencies in the candidate frequency set are sorted in ascending order of frequency. A greedy algorithm is then used to select frequencies one by one, ensuring that the frequency difference between each newly selected frequency and all previously selected frequencies is greater than or equal to the minimum interval constraint. The algorithm starts with the lowest frequency. If the frequency difference between the current frequency and the previously selected frequency satisfies the minimum interval constraint, it is added to the final set of first-order frequencies. Otherwise, it is skipped and the next frequency is checked. This approach ensures that the selected frequencies are evenly distributed across the frequency domain and do not interfere with each other.

[0078] The genetic algorithm outputs spectrum allocation results, including the dedicated sub-band range assigned to each target drone and the first frequency point set within that sub-band. The spectrum allocation results provide the dedicated spectrum resource foundation for the subsequent establishment of the master-slave architecture and the generation of frequency hopping sequences.

[0079] After optimization is complete, the ground station broadcasts the spectrum allocation results to all target drones via a control channel, including information such as the dedicated frequency band range, specific frequency points, and usage time for each drone. Upon receiving this allocation information, each drone configures its local frequency synthesizer and filter, preparing to communicate on the designated spectrum resources. Spectrum allocation information is protected using digital signature technology to prevent malicious tampering and forgery.

[0080] To ensure the dynamic adaptability of spectrum allocation, a spectrum allocation update mechanism has been established. Spectrum reallocation is automatically triggered when the spectrum environment changes significantly, the target drone mission priority is adjusted, communication quality degrades, or a new target drone is added. Significant changes in the spectrum environment refer to changes in the total spectrum range exceeding 20%, changes in the signal quality index exceeding 3 decibels, or an increase in interference power exceeding 10 decibels due to a new interference source. A target drone mission priority adjustment refers to a change in the mission priority value exceeding 2 levels or a change in the number of high-priority tasks exceeding 30%. Degraded communication quality refers to a bit error rate exceeding 10 to the power of -4, a decrease in the signal-to-noise ratio exceeding 5 decibels, or more than three communication interruptions per hour. The addition of a new target drone refers to the addition of a new target drone to the system, requiring dedicated spectrum resources and re-optimization of the overall spectrum allocation plan. The update cycle is set to 1-5 minutes, balancing the requirements of adaptability and stability. During reallocation, current communications are maintained, and a gradual handover method is used to gradually migrate communications to the new spectrum resources.

[0081] This intelligent spectrum management process achieves optimal spectrum resource allocation, providing a high-quality, dedicated spectrum resource foundation for subsequent master-slave architecture establishment and coordinated communications. Sub-band allocation and the determination of the first frequency point set ensure the rationality and effectiveness of resource allocation, meeting the complex requirements of dynamic communications for multiple drones.

[0082] Step 103: Calculate the communication quality index based on the power data, select the master target based on the task priority and the communication quality index, and establish a timing benchmark.

[0083] Specific methods include:

[0084] A comprehensive scoring method is used to select the primary target drone, based on two core factors: mission priority and communication quality. The communication quality index is calculated by comprehensively considering three key parameters: the signal-to-noise ratio reflects the fundamental quality of the communication link, the distance factor accounts for the impact of propagation losses, and the bit error rate reflects the actual communication performance. The signal-to-noise ratio is used as the positive factor in the quality index, while the distance factor multiplied by the bit error rate serves as the negative factor in the numerator. In the denominator, a ratio calculation is performed to obtain a comprehensive communication quality evaluation. The signal-to-noise ratio refers to the ratio of useful signal power to noise power, with units of decibels and a typical value range of 10-40 decibels. The calculation method is that the signal-to-noise ratio is equal to ten times the signal power divided by the common logarithm of the noise power; the distance factor refers to the free-space propagation loss caused by the distance from the target aircraft to the ground station, with units of decibels. The loss is calculated as 32.45 plus 20 times the common logarithm of the communication frequency plus 20 times the common logarithm of the distance, where the frequency is in MHz and the distance is in km. The common logarithm refers to the logarithm with base 10; the bit error rate refers to the ratio of the number of received erroneous bits to the total number of transmitted bits. The typical value range is 10 to the negative power of 6 to 10 to the negative power of 3. It is detected by cyclic redundancy check or forward error correction coding and obtained through observation statistics during real-time communication.

[0085] The weighting of the election algorithm reflects the principle of task orientation. The weight of task priority is typically set at 0.6, reflecting the importance of task execution. The weight of communication quality is set at 0.4, ensuring that the selected primary drone has good communication capabilities. A weighted sum is used to calculate the overall score for each drone, and the drone with the highest overall score is selected to serve as the primary drone. The advantage of using a comprehensive scoring method is that it balances task importance and communication capabilities, avoiding selection bias caused by a single metric. The comprehensive score is calculated as follows: 0.6 multiplied by the normalized task priority plus 0.4 multiplied by the normalized communication quality metric. Normalization involves mapping the original value to a range of 0 to 1. This is calculated by subtracting the minimum value from the original value and then dividing it by the maximum value minus the minimum value.

[0086] The election utilizes a distributed voting mechanism, with each drone voting independently based on its own observations to avoid single points of failure. The ground station serves as the central node for vote aggregation and result announcement, collecting all voting information and calculating the final results. To prevent malicious interference, digital signature technology is used in communication to verify the authenticity and integrity of votes. The distributed voting mechanism means that each node independently makes decisions and votes, eliminating the single point of failure associated with centralized decision-making. Digital signature technology uses RSA or the Elliptic Curve Digital Signature Algorithm (ECDSA) with a key length of 2048-4096 bits to ensure the unforgeability and non-repudiation of votes. The master drone's term is set at 5-10 minutes; significant performance degradation during this term can trigger a re-election early.

[0087] After the primary target drone is selected, it immediately collaborates with the ground station to establish a timing baseline. The timing baseline is a standardized time reference signal used to ensure strict time alignment among all target drones. Specifically, it is a composite signal with multiple frequency modulations. This timing baseline signal has the following clear technical characteristics: signal frequency accuracy of the order of 10 to the power of -12, time synchronization accuracy controlled within ±100 microseconds, signal coverage of at least 100 kilometers, and an anti-interference margin greater than 20 decibels. Signal frequency accuracy refers to frequency stability, using a cesium or rubidium atomic clock as a reference, with short-term stability ranging from 10 to the power of -11 to 10 to the power of -12. Time synchronization accuracy refers to the deviation of each node's clock from the standard time, achieved through two-way time transfer and propagation delay compensation. Signal coverage refers to the effective propagation distance of the timing reference signal, which is affected by transmit power, antenna gain, and propagation environment. Anti-interference margin refers to the ability to operate normally in an interference environment, improved through spread spectrum technology and error correction coding.

[0088] The steps to obtain a timing benchmark include:

[0089] The ground station obtains high-precision GPS or BeiDou time through a time and frequency receiver. This is the fundamental source of the entire timing reference. It is important to note that there is a fixed offset between GPS time and BeiDou time (BDT). The same time standard should be used in system design. The corresponding time and frequency receiver is a professional GPS or BeiDou receiving device.

[0090] The digital signal generator generates multiple standard frequency components of different frequencies based on the GPS or Beidou time obtained in the first step, including international standard frequencies such as 10MHz, 1MHz, and 100kHz. The corresponding digital signal generator uses direct digital frequency synthesis technology;

[0091] The signal processing unit uses a digital signal processor to perform frequency domain superposition synthesis on each standard frequency component, that is, digitally superimposing multiple standard frequency components such as 10MHz, 1MHz, and 100kHz to generate the final timing reference signal. The corresponding signal processing unit uses a high-performance DSP chip or FPGA, with a computing speed of up to 100 billion floating-point operations per second.

[0092] The timing reference signal is broadcast to all target drones through a high-power transmitter to ensure that every target drone within the coverage area can receive it.

[0093] The timing reference uses a composite signal technology called multi-frequency modulation. By simultaneously modulating multiple sinusoidal signals of different frequencies on a single carrier, a reference signal with high-precision time characteristics is formed. The advantage of using multi-frequency modulation technology is its strong anti-interference ability. Even if some frequency components are interfered with, the accuracy of the time reference can still be guaranteed for other components. Multi-frequency modulation refers to the simultaneous modulation of multiple baseband signals of different frequencies on a single carrier signal. The mathematical expression is that the modulated signal is equal to the carrier modulus multiplied by a cosine function. The parameter of the cosine function is twice pi times the carrier frequency times the time plus the sum of the modulation index of each modulated component times the second cosine function. The parameter of the second cosine function is twice pi times the frequency of each modulated component times the time.

[0094] The composition of the timing reference signal is based on a unified time system (GPS time or Beidou time), on which multiple modulation frequency components are superimposed. Each component has specific time offset, modulation frequency and initial phase angle parameters. The time offset is set to microsecond accuracy, the modulation frequency selects different international standard frequencies, and the initial phase angle is generated by a pseudo-random sequence to enhance the signal's anti-interference ability. The specific timing reference signal can be expressed as follows: the reference time is equal to the unified time plus the weighted combination of each modulation component, where the weight coefficient is determined by an optimization algorithm to ensure the stability and anti-interference of the signal. The time offset is the time difference set to distinguish different modulation components, with an accuracy of 1 microsecond; the international standard frequencies include 1MHz, 5MHz, 10MHz, etc., which are stipulated by the International Telecommunication Union; the pseudo-random sequence uses an m sequence or a Gold sequence, and the period length is , where n is the sequence order.

[0095] The digital signal generator utilizes a high-stability crystal oscillator and GPS or Beidou servicing technology to ensure frequency accuracy meets requirements. Digital signal generation utilizes direct digital frequency synthesis (DDS) technology, supporting arbitrary waveform generation and precise frequency control. Transmit power is adaptively adjusted based on propagation distance and environmental conditions to ensure that all slave drones receive a signal strength sufficient for synchronization. The slave drones are considered to be all other slave drones, excluding the master drone. High-stability crystal oscillators are temperature-compensated or oven-controlled crystal oscillators. GPS or Beidou servicing technology utilizes satellite signals to correct the frequency of the local oscillator. Direct digital frequency synthesis (DDS) directly generates the desired frequency signal digitally, achieving frequency resolution down to the microhertz level.

[0096] The digital signal generation device consists of five main components: a GPS or Beidou receiver module, a high-stability reference oscillator, a digital signal processor, a power amplifier, and an antenna system. The GPS or Beidou receiver module receives satellite signals and provides a standard time base. The high-stability reference oscillator maintains the local clock in the event of GPS or Beidou signal loss. The digital signal processor performs multi-frequency modulation signal synthesis operations. The power amplifier ensures sufficient signal transmission power. The antenna system is responsible for signal transmission and coverage. The GPS or Beidou receiver module includes an RF front-end, a baseband processor, and a clock output interface, capable of simultaneously tracking 8-12 satellites. The high-stability reference oscillator utilizes dual-furnace constant temperature control. The digital signal processor utilizes a multi-core ARM or DSP architecture to support parallel processing. The power amplifier utilizes a linear amplifier. The antenna system utilizes either an omnidirectional antenna or a directional antenna array.

[0097] The transmitted timing reference signal uses spread spectrum technology to enhance anti-interference capabilities and includes a periodic synchronization identification code to facilitate rapid synchronization and phase lock at the receiving end. The signal transmission power is adaptively adjusted based on the propagation distance and environmental conditions to ensure that all slave drones receive a signal strength that meets the synchronization requirements. Signal propagation adopts a line-of-sight mode, taking into account the effects of the earth's curvature and atmospheric refraction on the propagation path. Spread spectrum technologies include direct sequence spread spectrum and frequency hopping spread spectrum, with a processing gain of 10-30 decibels. The synchronization identification code is a specific digital sequence used for frame synchronization and clock recovery, typically using a Barker code or m-sequence. Adaptive power adjustment is performed based on the signal strength indication fed back by the receiving end, with an adjustment step size of 1 decibel. Line-of-sight propagation refers to the mode in which electromagnetic waves propagate in a straight line, and the propagation distance is limited by the antenna height and the curvature of the earth.

[0098] Each slave drone receives a timing reference signal and achieves precise synchronization with the time reference through an adaptive broadband phase-locked loop (PLL) structure. The PLL device consists of three core components: a digital phase detector, an adaptive loop filter, and a high-performance voltage-controlled oscillator. The digital phase detector is responsible for detecting the phase-angle difference and frequency offset between the local clock and the reference signal. The adaptive loop filter adjusts the filter parameters in real time according to the dynamic characteristics of the channel. The high-performance voltage-controlled oscillator provides a wide tuning range and fast response capability. The adaptive broadband PLL is a tracking system designed specifically for high-dynamic environments, capable of quickly tracking rapid changes in frequency and phase angle. The digital phase detector is implemented using a complex multiplier and CORDIC algorithm. The adaptive loop filter adopts a variable parameter design and automatically adjusts the bandwidth according to the Doppler frequency shift rate. The high-performance voltage-controlled oscillator uses direct digital frequency synthesis technology.

[0099] After receiving the timing reference signal from the target drone, it first undergoes signal preprocessing through a bandpass filter and automatic gain control circuit. The RF signal is then converted to a digital baseband signal via a digital downconverter. The adaptive phase-locked loop system includes a Doppler shift estimation module that calculates and compensates for frequency offsets caused by high-speed motion in real time. A digital phase detector compares the compensated reference signal with the local clock signal in terms of both angle and frequency to generate an error control signal. The adaptive loop filter dynamically adjusts the loop bandwidth based on the current motion state and channel variations, achieving an optimal balance between tracking speed and noise suppression. A high-performance voltage-controlled oscillator rapidly adjusts the frequency and angle of the local clock based on the filtered control signal, ultimately achieving high-precision synchronization with the reference signal. Doppler shift estimation utilizes a maximum likelihood estimation algorithm. The adaptive bandwidth adjustment algorithm determines the loop bandwidth based on the velocity change rate. The adaptive bandwidth is calculated as the base bandwidth plus the velocity change rate multiplied by an adjustment factor. The base bandwidth is 50 Hz, and the adjustment factor is determined through Kalman filtering optimization.

[0100] Employing an adaptive wideband phase-locked loop (PLL) design, the loop bandwidth dynamically adjusts based on the target drone's motion state, ensuring rapid tracking capabilities even under high-speed maneuvers. For high-speed target drone scenarios, the system integrates a Doppler pre-compensation algorithm that uses motion state information to predict Doppler shift and proactively pre-correct the frequency. Through multi-stage closed-loop feedback control, it automatically compensates for clock offsets caused by factors such as propagation delay variations, the Doppler effect, temperature drift, and oscillator aging. Achieving synchronization accuracy relies on intelligent optimization of adaptive control parameters, including real-time adjustment of the dynamic loop bandwidth, variable time constant, and adaptive damping coefficient. After multiple compensations and adaptive control, the deviation between the slave drone's local clock and the master timing reference remains within a preset deviation threshold, even under high-speed maneuvers, meeting the stringent timing requirements of coordinated communications. The dynamic loop bandwidth refers to the tracking bandwidth that is adjusted in real time according to the dynamic characteristics of the input signal. Narrow bandwidth is used in low-dynamic environments to reduce noise, and wide bandwidth is used in high-dynamic environments to increase tracking speed. Doppler pre-compensation predicts frequency offset through carrier motion parameters. Multi-level closed-loop control includes a coarse-tuning loop and a fine-tuning loop. The coarse-tuning loop is responsible for wide-range frequency capture, and the fine-tuning loop is responsible for high-precision angle tracking. The intelligent optimization of adaptive control parameters adopts a recursive least squares algorithm to adjust the control parameters in real time according to historical performance data.

[0101] Accurately calculating clock offset requires considering the combined effects of multiple dynamic factors. The calculation method includes: base offset is calculated as the target aircraft's local clock time minus the received synchronization time reference; propagation delay compensation is calculated based on the real-time distance and radio wave propagation speed, with the compensation value equal to the distance divided by the speed of light; Doppler effect compensation calculates the time deviation caused by frequency offset based on relative motion speed; and atmospheric propagation correction considers the impact of the ionosphere and troposphere on signal propagation speed. The final clock offset is equal to the base offset minus propagation delay compensation, minus Doppler effect compensation, and then minus atmospheric propagation correction. The real-time distance is obtained through two-way ranging technology; Doppler effect compensation is calculated as frequency offset divided by carrier frequency and multiplied by time interval; atmospheric propagation correction uses a standard atmospheric model with a correction accuracy of 0.1% of the propagation delay.

[0102] A multi-mode adaptive controller is used for precise clock offset correction. The controller features three operating modes: fast acquisition mode, precise tracking mode, and high-dynamic tracking mode. Fast acquisition mode is used for initial synchronization establishment, utilizing a large bandwidth for rapid convergence; precise tracking mode is used to maintain high-precision synchronization in stable environments; and high-dynamic tracking mode is specifically designed for high-speed maneuvers. The controller automatically selects the optimal operating mode based on the current error magnitude, rate of change, and predicted trend. This multi-mode control strategy significantly improves the system's adaptability to diverse motion states, ensuring ±1 millisecond synchronization accuracy under all conditions. The mode switching criteria are: fast acquisition mode is entered when the synchronization error is greater than 10 milliseconds; precise tracking mode is entered when the synchronization error is less than 1 millisecond and the rate of change is less than 0.1 milliseconds / second; and high-dynamic tracking mode is entered when the rate of change is greater than 10 meters per second squared. Control parameters are dynamically adjusted between modes: the loop bandwidth in fast acquisition mode is 1 kHz, the loop bandwidth in precise tracking mode is 50 Hz, and the loop bandwidth in high-dynamic tracking mode is adaptively adjusted based on dynamic parameters.

[0103] The overall performance of the master and target drone is evaluated in real time, and a master-slave switchover is automatically triggered when performance drops below a preset comprehensive threshold. Performance evaluation utilizes multi-dimensional metrics, including signal-to-noise ratio stability, communication reliability, and coverage area. Communication reliability is calculated through statistical analysis of bit error rate data; coverage is calculated by comparing the area of ​​the covered area with the total area required for coverage. A weighted summation of these three metrics is used to calculate a comprehensive performance evaluation, with a weight of 0.4 for signal-to-noise ratio stability, 0.4 for communication reliability, and 0.2 for coverage area. The overall performance evaluation ranges from 0 to 1. Among them, the signal-to-noise ratio stability refers to the amplitude of the change in the signal-to-noise ratio, which is measured by calculating the standard deviation of the signal-to-noise ratio within a certain time window. When the standard deviation is less than 2 decibels, it is evaluated as 1, and when it is greater than 10 decibels, it is evaluated as 0; the communication reliability refers to the stability of the communication link, which is evaluated by counting the number and duration of communication interruptions. When the bit error rate is less than 10 to the power of -5 and there is no interruption, it is evaluated as 1, and when the bit error rate is greater than 10 to the power of -3 or the interruption time exceeds 5 seconds, it is evaluated as 0; the coverage area refers to the geographical area where the main target aircraft can communicate effectively. It is affected by the transmission power, antenna radiation pattern and terrain. When the coverage rate reaches more than 95%, it is evaluated as 1, and when it is less than 60%, it is evaluated as 0.

[0104] Performance monitoring conducts a comprehensive assessment every 30 seconds, employing a multi-level threshold alarm mechanism. When the overall performance evaluation falls below the warning threshold of 0.7, an early warning signal is issued. When it falls below the danger threshold of 0.5, a master-slave switchover process is automatically triggered. The default value of the pre-set overall threshold is 0.5. Seamless handover technology is used during this process, with the pre-synchronized backup master drone immediately taking over, ensuring uninterrupted communication services.

[0105] After establishing stable timing synchronization, in order to obtain accurate channel state information for subsequent frequency hopping sequence generation and optimization, the communication method needs to perform high-precision time division multiplexing channel detection.

[0106] Step 104: divide the time axis into N+1 non-overlapping time slots based on the timing reference, and each target machine sends an orthogonal detection sequence in the corresponding time slot to obtain channel state information.

[0107] Specific as Figure 2 Shown, including:

[0108] Timeline unification and synchronization: All target drones and ground stations are strictly synchronized based on the timing reference established in step 103 to ensure that all nodes in the system have a completely consistent understanding of the timeline. The timing reference provides a unified time reference to achieve microsecond-level time synchronization.

[0109] The frame structure design, based on a unified timing benchmark, divides the continuous time axis into periodic time frames. The length of each time frame is determined by the number of target drones and the detection accuracy requirements, with a typical value of 10-100 milliseconds. The start time of the time frame is strictly determined according to the timing benchmark to ensure that the frame boundaries of all participating nodes are fully aligned.

[0110] Time slot division and allocation: Each time frame is divided into N+1 equal-length time slots. The time slot length is determined by dividing the time frame length by the total number of time slots. The first N time slots are allocated to N target drones for dedicated channel detection, and the N+1th time slot is allocated to the ground station for data aggregation and processing. Time slot allocation is strictly based on the timing benchmark. The start time of the time slot of the i-th target drone is equal to the target drone serial number minus one multiplied by the length of the single time slot, ensuring that the time slots are strictly arranged in time sequence and do not overlap.

[0111] A time slot guarding mechanism is established based on the timing benchmark to monitor in real time whether each target machine is sending probe signals strictly according to the allocated time slot. Illegal use of time slots will be immediately detected and corrected to ensure the strict implementation of time division multiplexing.

[0112] A time slot allocation table is established and broadcast to all participating nodes via dedicated control signaling. This table contains key information such as the precise start time, duration, and assigned target drone number for each time slot. The communication method implements a strict time slot guarding mechanism, monitoring in real time whether each target drone is accurately transmitting its probe signal according to its assigned time slot to prevent time slot conflicts and illegal occupation.

[0113] Target machines that violate time slot usage regulations will be temporarily disabled from the communication method until the issue is resolved and re-verified before normal participation can resume. Time slot usage statistics are recorded in real time, providing a basis for communication method optimization and fault diagnosis.

[0114] Each target drone transmits an orthogonal sounding sequence within its assigned dedicated time slot. The ground station calculates the detailed channel state of each link based on the received sounding signal, obtaining channel state information. The target drone transmits a known orthogonal sounding sequence, and the sounding signal received by the ground station is the signal after the orthogonal sounding sequence has propagated through the wireless channel—that is, the result of convolution of the original sounding sequence with the channel impulse response. The ground station extracts channel characteristic information by comparing the known orthogonal sounding sequence with the actual received sounding signal. Channel state information includes the channel impulse response, the number of multipath taps, and multipath propagation parameters. The channel impulse response is composed of the complex gain coefficients of multiple multipath components. Multipath propagation parameters include the time delay, modulus attenuation, and amplitude / angle offset of each path.

[0115] The propagation and reception process of the detection signal specifically includes: the orthogonal detection sequence sent by the target machine has known time domain and frequency domain characteristics; when the detection sequence propagates in the wireless channel, it will be affected by factors such as multipath propagation, fading, noise and interference; the detection signal received by the ground station is the result of the convolution operation of the original detection sequence and the channel impulse response and then superimposed with noise, that is, the received signal is equal to the response of the orthogonal detection sequence through the channel propagation plus the environmental noise; the received signal contains channel characteristic information, but is no longer the original orthogonal detection sequence, but a signal after channel distortion; the signal processor of the ground station uses the reverse engineering method to extract the channel impulse response through the known relationship between the orthogonal detection sequence and the received signal.

[0116] The specific method for obtaining the channel impulse response includes: realizing channel estimation by performing matched filtering and correlation processing on the received signal, firstly performing a correlation operation on the received detection signal and the known orthogonal detection sequence, and the correlation operation is performed by integrating the conjugate of the detection signal and the orthogonal detection sequence at different time delay positions, and the integration result is the channel impulse response estimate corresponding to the time delay position; due to the existence of multipath effects in the wireless propagation environment, the transmitted signal will reach the receiving end with different time delays through different paths, so the result of the correlation operation is manifested in the time domain as peaks at multiple time delay positions, each multipath component corresponds to a peak output by the correlation function, and the time position of the peak reflects the propagation delay of the path, and the complex value of the peak is the complex gain coefficient of the path, which contains modulus information and phase angle information. A peak refers to the local maximum point of the correlation function output in the time domain, specifically the peak point reached at a specific time delay. The amplitude and position of this peak point contain key information about multipath propagation: the peak position corresponds to the propagation delay of the multipath component, and the peak complex value contains the complex gain coefficient of the multipath component. The peak is formed when the received signal and the orthogonal detection sequence are aligned, and the correlation function output reaches its maximum value, forming a peak. When the delay is not aligned, the correlation function output is smaller, resulting in a distinct peak characteristic on the time domain correlation function curve. The number of peaks is equal to the number of valid multipath components, and each peak corresponds to a propagation path. The peak height reflects the signal strength of that path, and the peak width is affected by the autocorrelation characteristics of the detection sequence. The number of multipath taps is determined using a power threshold detection method. Multipath components exceeding a preset threshold are considered valid taps. The preset threshold is set at minus 20 decibels relative to the strongest path power. The number of taps is typically controlled between eight and sixteen to balance estimation accuracy and computational complexity. The specific number is dynamically adjusted based on the complexity of the propagation environment. The strongest path refers to the path with the highest power among all propagation paths in a multipath propagation environment.

[0117] The specific method for obtaining multipath propagation parameters includes: the time delay of each path is determined by the position of the peak, and the time delay is equal to the time position of the peak minus the arrival time of the direct path. The measurement accuracy reaches the nanosecond level, and the interpolation algorithm is used to improve the resolution of the delay estimation; the modulus attenuation is calculated by the modulus of the peak, and the attenuation value is equal to twenty times the modulus of the received signal divided by the common logarithm of the modulus of the transmitted signal, in decibels, with a typical value range of 0-40 decibels; the amplitude offset is obtained by calculating the amplitude of the peak, and the amplitude offset is equal to the amplitude of the received signal minus the amplitude of the transmitted signal, ranging from -π to π radians, and the amplitude measurement accuracy is 0.01 radians. The peak position refers to the coordinates of the peak on the time delay axis, determined by a peak detection algorithm. The detection algorithm first differentiates the correlation function to find the zero point, and then uses the second-order derivative to determine whether it is a peak. The modulus of the peak refers to the amplitude of the complex number of the peak, calculated by taking the square root of the sum of the squares of the real part and the imaginary part of the complex number. The amplitude of the peak refers to the phase angle of the complex number of the peak, calculated by calculating the ratio of the imaginary part to the real part of the complex number using the inverse tangent function, and the range is limited to -π to π. The accuracy of peak detection directly affects the accuracy of multipath parameter estimation. Multiple threshold detection and noise suppression algorithms are used to ensure the reliability of peak detection, while a super-resolution algorithm is used to separate similar peaks. The measurement of these parameters takes into account the influence of the Doppler effect and compensates for it using the carrier motion state information.

[0118] An orthogonal sounding sequence is a set of mutually orthogonal signal sequences. Its mathematical characteristic is that the cross-correlation function of any two different sequences is zero at zero time delay, and the autocorrelation function has a distinct peak at zero time delay. This orthogonality enables the receiver to effectively separate and identify signals from different transmitters, even if these signals arrive at the receiver at the same time. Orthogonal sounding sequences are primarily generated through pseudorandom code generation. Commonly used sequences include Gold sequences, Kasami sequences, and Walsh codes. Gold sequences are generated by modulo-2 addition of two m sequences, with a sequence length of 2 to the power of n-1, making them suitable for medium-sized systems. Kasami sequences are generated by extracting and combining m sequences, have a small maximum cross-correlation value, and are suitable for systems requiring strict orthogonality. Walsh codes are generated based on Hadamard matrices, with a sequence length of 2 to the power of n, making them suitable for synchronous systems. In practical applications, the selection of sequences must take into account factors such as system scale, anti-interference requirements, and computational complexity.

[0119] The channel state is mathematically described using the finite impulse response (FIR) model, which treats a complex wireless channel as a linear, time-invariant system consisting of multiple propagation paths with varying delays. The channel impulse response is a time-domain representation of channel characteristics, while the FIR model is a specific implementation that imposes a finite memory limit on the system, ensuring that the system responds only to inputs within a finite timeframe.

[0120] This communication method uses a correlation receiver architecture to accurately estimate the channel impulse response. The process begins by transmitting a known sounding sequence. Then, by calculating the correlation between the sounding sequence and the actual received signal, the channel impulse response characteristics are extracted from the correlation result. In practical implementation, the correlator employs fast Fourier transform (FFT) technology to convert time-domain convolution into frequency-domain multiplication, significantly improving computational efficiency and real-time processing capabilities.

[0121] In complex multipath environments, traditional correlation methods have limitations in delay resolution. Therefore, super-resolution algorithms (such as MUSIC or ESPRIT) are employed to accurately estimate multipath components. These algorithms can identify and separate multipath components with very close delays, surpassing the resolution limitations of conventional Fourier analysis and providing higher estimation accuracy than traditional methods. Complete multipath parameter estimation results include information such as delay, modulus attenuation, and amplitude offset for each path. This information constitutes accurate channel impulse response parameters, providing a reliable foundation for subsequent frequency hopping sequence generation and optimization.

[0122] The complete channel estimation steps include: each target machine sends a known orthogonal detection sequence strictly according to the allocated time slot to ensure the time accuracy and sequence orthogonality of signal transmission; the ground station receives all detection signals and performs precise time slot separation processing to extract the signals belonging to different target machines respectively; the signal in each time slot is specially correlated, and the known detection sequence is used as a matched filter to extract the channel impulse response of the link from the received signal; the channel impulse response is optimized through the minimum mean square error estimation method to reduce the impact of noise and interference on the estimation accuracy; all channel estimation results are statistically analyzed and quality evaluated, and then stored in the channel status database to obtain channel status information, which is promptly broadcast to the relevant target machine nodes.

[0123] The channel estimation processor is implemented using a dedicated digital signal processing chip or field-programmable gate array to ensure that it meets the stringent requirements of real-time processing. The communication method requires that the normalized mean square error of the channel estimate be less than 0.01. This threshold, determined through extensive simulation experiments, ensures the accuracy of subsequent frequency hopping sequence generation. Estimation accuracy is continuously verified using a dedicated pilot signal, which uses QPSK modulation with a power of -10dBm and is transmitted every 100 milliseconds for a duration of 1 millisecond. A complete channel state history database is established, using a time series storage structure to store the last 24 hours of channel state data. With a storage capacity of 10GB, this database provides reliable data support for subsequent frequency hopping sequence generation, conflict prediction, and communication method optimization.

[0124] After obtaining high-precision channel state information, combined with spectrum allocation results and timing benchmarks, the communication method has all the conditions to generate an adaptive frequency hopping sequence.

[0125] Step 105: Convert the channel state information into a frequency hopping key; and generate a frequency hopping sequence based on the frequency hopping key and the spectrum allocation result.

[0126] Specific as Figure 3 and Figure 4 Shown, including:

[0127] Frequency hopping key generation uses a multi-channel feature fusion method to improve the randomness and unpredictability of the key. Key information is extracted from the channel state information obtained in step 104 for key generation. The key information includes the modulus of the complex gain coefficient, the amplitude of the complex gain coefficient, and the time rate of change of the channel state. The modulus of the complex gain coefficient reflects the amplitude fading characteristics of the channel; the amplitude of the complex gain coefficient reflects the phase variation characteristics of the channel; and the time rate of change of the channel state reflects the dynamic variation characteristics of the channel.

[0128] The complex gain coefficient specifically refers to the complex gain coefficient of the strongest multipath component in the channel impulse response. The strongest multipath component is selected by power-ranking all multipath components in the channel impulse response obtained in step 104 and selecting the component with the largest modulus of the complex gain coefficient as the strongest multipath component. This component typically corresponds to a direct path or a primary reflection path and has the highest signal-to-noise ratio and the most stable estimation accuracy.

[0129] The time domain channel impulse response is obtained by the correlation receiver processing in step 104. The channel impulse response is composed of multiple complex gain coefficients, each of which corresponds to a multipath component. The complex gain coefficient is in complex form and consists of two components, a real part and an imaginary part. The real part reflects the in-phase component, and the imaginary part reflects the orthogonal component. The complex gain coefficient is calculated by a digital signal processor. The square of the real part and the square of the imaginary part of the complex gain coefficient are added, and then the square root of the result is taken to obtain the absolute value of the complex number, which is the modulus of the complex gain coefficient. The imaginary part of the complex gain coefficient is divided by the real part, and the inverse tangent function is applied to the result to obtain the phase angle of the complex number, which is the amplitude angle of the complex gain coefficient.

[0130] The time rate of change of the channel state is obtained by the ratio of the modulus of the change of the complex gain coefficient at adjacent moments to the time interval.

[0131] The specific method for obtaining the change in the complex gain coefficient at adjacent moments includes:

[0132] The complex gain coefficients of the strongest multipath components at the last 10 sampling moments are stored. The cache uses a first-in-first-out (FIFO) queue structure with a queue length of 10 complex elements, each of which occupies 8 bytes of storage space.

[0133] To reduce the influence of noise, a three-point moving average filter is performed on the original complex gain coefficient data. The filtering method is to multiply the complex gain coefficient at the previous moment by 0.25, the complex gain coefficient at the current moment by 0.5, and the complex gain coefficient at the next moment by 0.25, and then add the three together to obtain the filtered complex gain coefficient.

[0134] The change in complex gain coefficients between adjacent moments is calculated by subtracting the filtered complex gain coefficients from the previous moment, where the time interval is the sampling interval. The change calculation is performed in the complex domain and includes both the real and imaginary changes, and the result is also in complex form.

[0135] The key information is independently quantized, and the complex gain coefficient modulus is converted to an integer value in the range of 0 to 255 using an 8-bit quantizer. The modulus is multiplied by 255 and rounded down to obtain a digital representation of the modulus. The complex gain coefficient amplitude angle, ranging from negative π to positive π radians, is mapped to an integer value between 0 and 255. The amplitude angle is first added to π, then multiplied by 255, divided by 2π, and finally rounded down to obtain a digital representation of the amplitude angle. The time rate of change is logarithmically transformed and normalized before being quantized to 8 bits. The quantization method is to first add 1 to the time rate of change, take the base-10 logarithm, multiply by 255, divide by the logarithm of the historical maximum rate of change plus 1, and finally round down to obtain a digital representation of the time rate of change.

[0136] The three 8-bit quantization results are combined through a bit concatenation operation to generate a 24-bit base key, where the modulus quantization result is the upper 8 bits, the argument quantization result is the middle 8 bits, and the time rate quantization result is the lower 8 bits. These are concatenated in sequence to form the 24-bit base key. The modulus quantization result occupies bits 23 to 16, the argument quantization result occupies bits 15 to 8, and the time rate quantization result occupies bits 7 to 0.

[0137] To enhance the randomness and diffusion of the key, the 24-bit base key is shifted and XORed. The 24-bit base key is XORed with the result of right-shifting it 8 bits, and then XORed with the result of right-shifting it 16 bits to create a mixed key. This process causes each 8 bits of the original 24-bit key to influence each other, enhancing the random distribution of the key.

[0138] The mixed key is expanded to 128 bits through multiple bit shifts and concatenation operations. The specific structure is as follows: bits 127 to 104 are the original 24-bit value of the mixed key; bits 103 to 80 are the lower 24-bit value of the mixed key after shifting left by 8 bits; bits 79 to 56 are the lower 24-bit value of the mixed key after shifting left by 16 bits; bits 55 to 32 are the lower 24-bit value of the mixed key after shifting left by 24 bits; bits 31 to 8 are the lower 24-bit value of the mixed key after bitwise inversion; bits 7 to 0 are the lower 8-bit value padded with the mixed key.

[0139] The final generated 128-bit frequency hopping key is used in the subsequent frequency hopping sequence generation process to ensure that the key has sufficient length and good random properties.

[0140] The main frequency hopping sequence is a frequency sequence that changes with time and is expressed as ,in Indicates the The hopping frequency of the time slot, the sequence length The number of frequencies, typically 1,000 to 10,000, is determined based on the communication cycle and frequency hopping rate. The master hopping sequence is generated using a direct mapping method from key to frequency. The core algorithm of the mapping function performs a modulo operation on the size of the first frequency set using the generated hopping key. The remainder is added by one, and the resulting index is used to select the corresponding frequency from the pre-established first frequency set as the hopping frequency at the current moment. The complete master hopping sequence is generated by repeatedly applying this mapping method to consecutive time slots. The hopping key for each time slot is dynamically updated based on real-time channel state information, ensuring the sequence's adaptability and anti-predictability.

[0141] This mapping method ensures uniform distribution and good randomness in frequency selection. The mapping function is implemented using an efficient table lookup method. A key-to-frequency mapping table of 65536 x 4 bytes is pre-established and regularly updated, allowing the frequency allocation strategy to be adjusted promptly based on changes in the spectrum environment.

[0142] The method for generating a frequency hopping sequence of a slave target machine includes: the slave target machine generates a slave frequency hopping key based on the channel state information it receives, and the slave frequency hopping key is 128 bits long; the slave frequency hopping key is processed by the SHA256 hash function to enhance randomness and anti-prediction capabilities; the master frequency hopping sequence information broadcast by the master target machine is received, and the information includes parameters such as the current frequency hopping frequency, sequence index and timestamp; a specially designed orthogonal mapping function is applied to generate the frequency hopping sequence of the slave target machine in combination with the hashed slave frequency hopping key and the master frequency hopping sequence information; and the final frequency point selection and confirmation are performed within the dedicated sub-band allocated in step 102.

[0143] The SHA256 algorithm is used as the hash function for the slave machine's key. The SHA256 algorithm converts an input key of any length into a fixed 256-bit hash value through a series of complex bitwise operations, including circular left shifts, logical XOR, and modulo operations. The SHA256 algorithm was chosen for its cryptographic security and uniform output distribution, making it a widely recognized secure hash algorithm.

[0144] The specific implementation methods of the orthogonal mapping function include:

[0145] According to the spectrum allocation result of step 102, each slave drone obtains the corresponding dedicated sub-band frequency range, and the first frequency point set in the sub-band Expressed as ,in For the The first one from the target drone The first frequency point, For the The frequency points are arranged in ascending order to facilitate subsequent indexing operations.

[0146] The generation of the hopping frequency of the slave target machine adopts a four-element XOR mixed mapping method, which is specifically implemented by converting the four values ​​of the slave hopping key after hash processing of the slave target machine, the hopping frequency at the current moment in the main hopping sequence, the eight-bit left shift result of the slave target machine number and the current timestamp into a thirty-two-bit data format and performing an XOR operation to obtain a composite key value, and then performing a modulo operation on the composite key value and the number of the first frequency point of the slave target machine to obtain the basic frequency point index; since the array index starts from zero and the frequency point number starts from one, it is necessary to add one to the basic index to obtain the final frequency point selection index; finally, the frequency point at the corresponding position is selected from the first frequency point set of the corresponding slave target machine through the index value as the hopping frequency at the current moment.

[0147] The slave number is a unique numerical identifier assigned to each slave drone. It is used to distinguish different slave drones in the orthogonal mapping function and generate differentiated frequency hopping sequences. Slave numbers are eight-bit unsigned integers ranging from 1 to 255. Identifiers are assigned sequentially: the first slave drone to join the system is assigned identifier 1, the second is assigned identifier 2, and so on, ensuring that each slave drone has a unique digital identity. Identifiers are uniformly assigned by the ground station when a slave drone initially joins the system and transmitted via secure signaling. Upon receiving the identifier, the slave drone confirms it, establishing a binding between the identifier and the actual device. Once assigned, the identifier remains unchanged throughout the communication session. When a slave drone exits the system, its identifier is recycled and can be assigned to a new slave drone.

[0148] To ensure the correct execution of the XOR operation, the four values ​​involved in the operation need to be uniformly converted to a 32-bit unsigned integer format. The specific conversion method includes: directly taking the lower 32 bits of the SHA256 algorithm output result from the hashed slave hopping key as the first operand; the hopping frequency at the current moment in the main hopping sequence is expanded to 32 bits by truncating or padding the integer part of the frequency value as the second operand; first performing an eight-bit left shift operation on the slave target machine number to obtain a 16-bit value, and then padding the high bit with zeros to expand it to 32 bits as the third operand; the current timestamp takes the lower 32 bits of the system time or converts it to 32 bits by truncating the high bit value as the fourth operand. After the four 32-bit values ​​are converted to a unified format, an XOR logical operation is performed bit by bit, and each bit is XORed separately, and finally a 32-bit composite key value is obtained for subsequent modulo operations and frequency index calculations.

[0149] The lower 32 bits of the hashed slave hopping key provide the underlying randomness. The current frequency of the master hopping sequence ensures correlation and synchronization with the master sequence. The left-shifted eight-bit slave number ensures the difference and independence of the mapping results between different slaves. The current timestamp adds variability and dynamic characteristics to the time dimension. These four parameters are fully mixed through a bit-by-bit XOR operation to ensure a statistically uniform distribution and high randomness in the mapping results.

[0150] Because each slave drone has an independent sub-band and a non-overlapping set of first frequencies, meaning no two slave drones have common frequencies, the frequency hopping frequencies generated by the orthogonal mapping function will never be the same at any given moment, ensuring complete orthogonality. The master drone uses a dedicated sub-band different from that of the slave drones, further ensuring orthogonality between the master and slave sequences.

[0151] After each mapping calculation is completed, the system automatically verifies the validity of the selected frequency, including whether it falls within the allocated sub-band, meets signal quality requirements, and conflicts with other target aircraft frequencies at the current moment. Verification is achieved through real-time table lookup and comparison operations. If an invalid frequency is found, the system automatically selects the next candidate frequency in the first frequency set, attempting up to three times to ensure the continuity and reliability of the frequency hopping sequence.

[0152] The frequency hopping cycle length is dynamically adjusted within a preset timeframe based on system load and real-time requirements, matching the synchronization accuracy of the timing reference to ensure timing consistency. When a sequence violating the orthogonality constraint is detected, the communication method immediately marks it invalid and triggers a regeneration process. A dedicated orthogonality statistics database is established to record the results of each orthogonality check, including information such as the check time, the number of sequences checked, and the number of conflicts found. This database continuously analyzes sequence quality and potential conflict probabilities, providing data support for communication method optimization.

[0153] A specially designed handshake protocol ensures precise synchronization of the frequency hopping sequences of all nodes. Sequence verification uses a cyclic redundancy check (CRC) method, performing a polynomial operation on the generated frequency hopping sequence to generate a 16-bit checksum. The receiving end uses the same calculation method to generate the checksum and compare it to verify the integrity and correctness of the received sequence.

[0154] The cyclic redundancy check polynomial uses the CRC-16-CCITT standard, with the generator polynomial being the sixteenth plus the twelfth plus the fifth plus one. This standard is widely used in communication systems and offers excellent error detection capabilities. The handshake protocol employs a three-way handshake mechanism to ensure the reliability of sequence synchronization. The handshake process consists of three phases: sequence request, sequence confirmation, and synchronization completion. If a checksum error is detected, the communication method automatically requests retransmission, with a maximum of three retries at a 100 millisecond interval. If failure persists, the sequence regeneration process is triggered.

[0155] After generating high-quality frequency hopping sequences, in order to cope with dynamic environmental changes and prevent potential conflicts, the communication method needs to establish a prediction-based dynamic sequence update and conflict avoidance mechanism.

[0156] Step 106, based on the channel state information and the frequency hopping sequence, which includes the master frequency hopping sequence and the slave frequency hopping sequence, a statistical method is used to calculate the frequency conflict probability at a future time, and a conflict probability threshold is set. When the frequency conflict probability exceeds the set conflict probability threshold, an adaptive adjustment mechanism is used to update the sequence; the occurrence of frequency hopping sequence conflicts is actively prevented to ensure the continuous and stable operation of the communication system in a dynamic environment.

[0157] Intelligent frequency conflict probability calculation calculates the probability of a frequency conflict occurring in the future based on channel state information and the generated frequency hopping sequence. This calculation employs statistical analysis to calculate the probability of frequency overlap in the future for all possible target-drone combinations. The summation then yields the overall system conflict probability.

[0158] The collision probability calculation utilizes the Monte Carlo statistical method, which estimates probability values ​​through a large number of random samplings, improving the accuracy and reliability of the probability estimation. The Monte Carlo sampling frequency is set to 10,000, determined based on statistical convergence analysis, with a sampling interval of 1 millisecond, covering a time window of 100 milliseconds into the future. The calculation frequency is set to 10 times per second, enabling real-time monitoring of collision risk. The communication method sets a collision probability threshold based on the communication service quality requirements, typically between 0.05 and 0.1. This threshold range is determined based on system reliability requirements.

[0159] When the predicted frequency collision probability exceeds a set collision probability threshold, the communication method automatically uses various adaptive adjustment mechanisms to adjust the sequence, reducing the collision risk to an acceptable level. Adaptive adjustment mechanisms are selected based on the principle of minimizing the impact on system performance, and include methods such as sequence offset adjustment, alternate sequence switching, and frequency point replacement.

[0160] Sequence offset adjustment is achieved by adding an appropriate frequency offset to the original frequency hopping sequence, while maintaining the basic structure of the sequence. The frequency offset is determined by the sub-band spacing in the spectrum allocation step 102 and is typically 10% of the sub-band bandwidth. For example, when the sub-band bandwidth is 10 MHz, the frequency offset is 1 MHz. Backup sequence switching uses pre-generated and verified alternative frequency hopping sequences to ensure fast response. The number of backup sequences is set to twice the number of primary frequency hopping sequences. The frequency point replacement method performs local adjustments on specific conflicting frequencies to minimize the impact on overall communication performance.

[0161] The adjustment strategy uses a hierarchical threshold judgment mechanism, selecting appropriate response measures based on the different levels of conflict probability. When the conflict probability is high, a pre-prepared backup sequence is prioritized for complete switching; when the conflict probability is medium, the sequence is partially modified and optimized; when the conflict probability is low but still exceeds the safety limit, the original sequence is maintained while increasing the monitoring frequency.

[0162] The communication method establishes two key probability thresholds: a high-conflict probability threshold of 0.1, exceeding which triggers an immediate switch to the backup sequence; and a low-conflict probability threshold of 0.05, serving as a safety margin for normal operation. These thresholds are determined based on extensive simulation experiments, which consist of multiple independent runs, and actual test data derived from actual flight tests. The adjustment strategy is implemented using an expert system, which makes intelligent decisions based on historical experience and operational rules.

[0163] All adjustments utilize a smooth transition mechanism to avoid interruptions in communication services. Backup sequences are pre-generated, verified, and stored in fast-access memory, ensuring immediate switchover when needed. The number of backup sequences is set to twice the number of primary hopping sequences, using the same generation algorithm as the primary hopping sequence but a different initial key. Adjustment commands are sent to the relevant target aircraft via reliable control signaling, protected by BCH coding with 3-bit error correction capability, and containing detailed switchover timing and new sequence parameters.

[0164] Through the complete implementation of this embodiment, an efficient and reliable coordinated dynamic frequency hopping communication system is established between multiple drones and the ground station. Each step is closely linked to form a complete closed-loop control system, effectively solving the multi-link conflict and synchronization problems.

[0165] Example 2:

[0166] See Figure 5 As shown, a dynamic spectrum hopping communication device for a target drone and a ground station is provided, which is used to store computer-readable instructions. When the computer-readable instructions are read, the dynamic spectrum hopping communication method for a target drone and a ground station can be executed. The device includes:

[0167] A collection module 701 collects status information, including task priority and total spectrum range;

[0168] Allocation module 702 monitors power data in real time and evaluates signal quality indicators based on the power data; based on the total spectrum range and the signal quality indicators, divides the total spectrum range into N sub-bands to obtain spectrum allocation results;

[0169] The timing module 703 calculates the communication quality index based on the power data, selects the master target based on the task priority and the communication quality index, and establishes a timing benchmark;

[0170] The channel module 704 divides the time axis into N+1 non-overlapping time slots based on the timing reference. Each target drone sends an orthogonal sounding sequence in the corresponding time slot to obtain channel state information.

[0171] The frequency hopping module 705 converts the channel state information into a frequency hopping key; and generates a frequency hopping sequence based on the frequency hopping key.

[0172] The above describes an embodiment of the present invention, but this embodiment is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Ordinary technicians in this field can also make more forms of equivalent embodiments based on the inspiration of this embodiment, all of which are protected by this embodiment.

Claims

1. A dynamic spectrum hopping communication method between a target drone and a ground station, characterized in that: The following steps are involved: Collect status information, including task priority and total spectrum range; Monitor power data in real time and evaluate signal quality indicators based on the power data. Based on the total spectrum range and signal quality indicators, divide the total spectrum range into N sub-bands to obtain spectrum allocation results. Calculate communication quality indicators based on power data, select the master drone based on task priority and communication quality indicators, and establish a timing benchmark based on GPS or Beidou time. Based on the timing reference, the time axis is divided into N+1 non-overlapping time slots. Each target machine sends an orthogonal sounding sequence in the corresponding time slot to obtain channel state information. The channel state information is converted into a frequency hopping key; and a frequency hopping sequence is generated based on the frequency hopping key and the spectrum allocation result.

2. The dynamic spectrum hopping communication method between a target drone and a ground station according to claim 1, wherein: The frequency hopping sequence includes a master frequency hopping sequence and a slave frequency hopping sequence. The steps for generating the frequency hopping sequence include: The master target machine uses the generated master frequency hopping key to perform a modulo operation on the size of the first frequency point set in the spectrum allocation result. The remainder obtained is added by one as the index value. The frequency point at the corresponding position in the first frequency point set is selected as the frequency hopping frequency at the current moment. The frequency hopping frequencies at multiple moments constitute the master frequency hopping sequence. The first frequency point set is the set of frequencies that meet the minimum interval constraint and the signal quality index is greater than the preset red threshold. The slave target machine generates a slave frequency hopping key according to the received channel state information; processes the slave frequency hopping key through the SHA256 hash function; receives the master frequency hopping sequence broadcast by the master target machine; and applies an orthogonal mapping function to combine the hashed slave frequency hopping key and the master frequency hopping sequence to generate a slave frequency hopping sequence.

3. The dynamic spectrum hopping communication method between a target drone and a ground station according to claim 2, wherein: The frequency hopping key includes a master frequency hopping key and a slave frequency hopping key. The frequency hopping key is generated by using a multi-channel feature fusion method to extract the key information of the path with the largest complex gain coefficient modulus in the channel state information. The key information includes the modulus of the complex gain coefficient, the argument of the complex gain coefficient, and the time rate of change of the channel state. Quantize the key information, concatenate the quantization results to obtain a basic key, shift and XOR the basic key to obtain a mixed key; perform multiple shift and concatenation operations on the mixed key to obtain a frequency hopping key; The modulus of the complex gain coefficient is the sum of the square of the real part and the square of the imaginary part of the complex gain coefficient, and then the square root of the sum is taken; the amplitude of the complex gain coefficient is the imaginary part of the complex gain coefficient divided by the real part, and the inverse tangent function of the result is obtained; the time rate of change of the channel state is obtained by the ratio of the modulus of the change in the complex gain coefficient at adjacent moments to the time interval, and the change in the complex gain coefficient at adjacent moments is obtained by subtracting the complex gain coefficient at the previous moment from the complex gain coefficient at the current moment.

4. The dynamic spectrum hopping communication method between a target drone and a ground station according to claim 2, wherein: The orthogonal mapping function converts the hashed slave frequency hopping key, the current frequency hopping frequency in the master frequency hopping sequence, the slave target machine number and the current timestamp into thirty-two bits and then performs an XOR operation to obtain a composite key value; a modulo operation is performed on the composite key value and the number of the first frequency points in the slave target machine, and the remainder obtained is added by one as the basic frequency point index; and the frequency point at the corresponding position is selected from the first frequency point set of the corresponding slave target machine through the basic frequency point index as the frequency hopping frequency at the current moment.

5. The dynamic spectrum hopping communication method between a target drone and a ground station according to claim 1, wherein: The specific steps of dividing the time slot into N+1 non-overlapping time slots include: All target drones and ground stations are time synchronized based on the timing reference; Split the continuous time axis into periodic time frames; Each time frame is divided into N+1 time slots of equal length. The first N time slots are allocated to N target drones, and the N+1th time slot is allocated to the ground station.

6. The dynamic spectrum hopping communication method between a target drone and a ground station according to claim 5, characterized in that: The sounding signal is the signal generated by the orthogonal sounding sequence after propagation through the wireless channel. The orthogonal sounding sequence is a signal sequence with mutually orthogonal characteristics. The channel state information includes the channel impulse response, the number of multipath taps, and the multipath propagation parameters. The channel impulse response is composed of the complex gain coefficients of multiple multipath components. The multipath propagation parameters include the time delay, modulus attenuation, and amplitude offset of each path. The detection signal is correlated with a known orthogonal detection sequence. The correlation operation is performed by integrating the conjugate of the detection signal and the orthogonal detection sequence at different time delay positions. The integration result is the channel impulse response corresponding to the corresponding time delay position. The number of multipath taps is determined by a power threshold detection method. Multipath components exceeding a preset threshold are determined to be valid taps. The number of valid taps is the number of multipath taps. The path delay is determined by the position of the channel impulse response in the time domain. The delay is equal to the time position of the peak minus the arrival time of the direct path. The modulus attenuation is calculated from the modulus of the peak value; The argument shift is obtained by calculating the argument of the peak.

7. The dynamic spectrum hopping communication method between a target drone and a ground station according to claim 1, wherein: The comprehensive score of each target aircraft is calculated by weighted summation, and the target aircraft with the highest comprehensive score is selected as the main target aircraft; the comprehensive score is the sum of the task priority item and the communication quality index item, where the task priority item is the normalized task priority multiplied by the preset task priority weight, the communication quality index item is the normalized communication quality index multiplied by the preset communication quality weight, and the communication quality index is the ratio of the signal-to-noise ratio to the bit error factor; among them, the bit error factor is the distance factor multiplied by the bit error rate, the signal-to-noise ratio is the ratio of the useful signal power to the noise power, the distance factor is the propagation loss caused by the distance from the target aircraft to the ground station, and the bit error rate is the ratio of the number of received error bits to the total number of transmitted bits.

8. The dynamic spectrum hopping communication method between a target drone and a ground station according to claim 1, wherein: The steps to establish a timing baseline include: Get GPS or BeiDou time from ground stations; Generate multiple standard frequency components of different frequencies according to GPS or Beidou time, including 10MHz, 1MHz, and 100kHz; Perform frequency domain superposition and synthesis of each standard frequency component to generate a timing reference signal; Broadcast the timing reference signal to all target machines.

9. The dynamic spectrum hopping communication method between a target drone and a ground station according to claim 1, wherein: The power data includes signal power, noise power and interference power. The signal quality index is the ratio of signal power to the total noise. The total noise is the sum of noise power and interference power. The spectrum allocation result includes the sub-band allocated to each target aircraft and the first frequency point set in the corresponding sub-band.

10. A dynamic spectrum hopping communication device between a target drone and a ground station, characterized in that: It is used to store computer instructions. When the computer instructions are read, the dynamic spectrum hopping communication method between the target drone and the ground station according to any one of claims 1 to 9 is executed. The device includes: The collection module collects status information, including task priority and total spectrum range; The allocation module monitors power data in real time and evaluates signal quality indicators based on the power data. Based on the total spectrum range and signal quality indicators, it divides the total spectrum range into N sub-bands to obtain the spectrum allocation results. The timing module calculates the communication quality index based on the power data, selects the master target based on the task priority and the communication quality index, and establishes the timing benchmark; The channel module divides the time axis into N+1 non-overlapping time slots based on the timing reference. Each target machine sends an orthogonal detection sequence in the corresponding time slot to obtain channel state information. The frequency hopping module converts the channel state information into a frequency hopping key and generates a frequency hopping sequence based on the frequency hopping key.

Images